AI Chatbot Regulation, State by State: A 2026 Tracker
A dated reference to state chatbot laws, their effective dates and enforcement. The broader survey was checked in October 2026; the November 10 update adds New York’s October 28 signing of S9051B, whose substantive requirements start January 1, 2027. Historical bill counts below come from FPF’s September report.
Why this is a patchwork
AI chatbot regulation is being written state by state, and quickly. The laws share a common core but differ in who they cover, when they start and how they're enforced. This page is the reference version: what's in force, what's signed but not yet in force, what's still waiting, and where enforcement is coming from lawsuits rather than new statutes. For the lawsuits against individual companies, see our AI chatbot lawsuits tracker and Character.AI timeline.
Laws in effect at the October 2026 review
- California — SB 243 (in effect since January 1, 2026). Companion chatbot operators must disclose that the chatbot is AI, follow a protocol for suicidal ideation and self-harm, and add protections for known minors, including break reminders and limits on sexual content. People harmed can sue for the greater of actual damages or $1,000 per violation. See our SB 243 explainer.
- New York — AI companion models law (General Business Law Article 47, in effect since November 5, 2025). Requires AI disclosure and a protocol to detect and respond to suicidal ideation and self-harm; enforced by the attorney general.
- Illinois — Public Act 104-0054 (HB 1806, the WOPR Act, 2025). Narrower: it bars AI from providing therapy or presenting itself as a licensed mental-health provider, rather than regulating companion chatbots in general.
- Utah — HB 452 (signed March 25, 2025). Covers mental-health chatbots: AI disclosure, limits on selling or sharing users' health information, and limits on advertising, including a ban on targeting ads based on what users type. The Division of Consumer Protection can fine up to $2,500 per violation (summary).
Signed in 2026, taking effect later
- California — SB 1119, "Adam's Law" (signed September 10, 2026; most duties start July 1, 2027). Builds on SB 243 for child users: annual child-safety risk assessments, independent audits, parent-adjustable default protections and time limits, crisis protocols, and limits on advertising to children and selling their data. Public prosecutors can seek up to $5,000 per affected child per negligent violation and $15,000 per intentional violation, and children who suffer actual harm can sue (FPF report).
- Oregon — SB 1546 (effective January 1, 2027). AI disclosure, extra protections for minors, suicide-risk protocols with 988 referrals, and annual reporting. Private right of action for the greater of actual damages or $1,000 per violation (Mayer Brown).
- Washington — HB 2225 (effective January 1, 2027). AI disclosure at the start of a conversation and every three hours (every hour for minors), a ban on sexually explicit content and manipulative engagement tactics for minors, and self-harm protocols. Enforced under the state Consumer Protection Act, which lets consumers sue (Mayer Brown).
- Nebraska — LB 525, the Conversational AI Safety Act (approved April 14, 2026). Safety and transparency duties for conversational AI; Idaho's SB 1297 follows a similar model (Orrick).
- Connecticut — SB 5. Age verification for minors, a ban on chatbots creating false social or emotional connections, and suicidal-ideation protocols (CDT summary).
- Others signed in 2026, according to the FPF tracker and CDT, include laws in Colorado, Georgia, Hawaii, Iowa, Maine, New Hampshire, Rhode Island and Tennessee. New York’s later S9051B signing is addressed separately below. FPF's September report counts 13 chatbot laws signed in 2026 and 18 across 2025 and 2026. Their scope varies widely, from AI-therapy bans to broader chatbot safety rules, so check each text before relying on it.
New York S9051B: signed, requirements start in 2027
The Senate’s official action record lists S9051B as signed on October 28, 2026, Chapter 368. Section 3 sets January 1, 2027 as the effective date, with rulemaking authorized beforehand. The law restricts unsafe AI companion features for minors and requires age assurance for adult access, subject to its scope and exceptions. It authorizes attorney-general enforcement, including civil penalties up to $25,000 per violation. It is no longer awaiting a signature; the substantive requirements were not yet in force on November 10.
Enforcement through lawsuits, not new statutes
States don't need a chatbot law to act. Attorneys general and other agencies are using existing consumer-protection and licensing laws:
- Kentucky sued Character.AI on January 8, 2026 under its consumer-protection and consumer-data laws, alleging harm to minors.
- Pennsylvania sued Character.AI on May 5, 2026, alleging chatbots posed as licensed medical professionals.
- Texas opened an investigation into Meta AI Studio and Character.AI in August 2025 over mental-health claims, and sued Snap in February 2026.
- Utah and New Mexico are suing Snap; Utah's case centres on the My AI chatbot. Details are in our lawsuits tracker.
The practical point: a state with no chatbot-specific law can still bring a case over chatbot conduct.
The common core
Across enacted laws and pending bills, the same requirements keep appearing:
- Telling users they are talking to an AI, often repeatedly during long sessions.
- Extra protections for minors: content limits, reminders, parental tools, age assurance, or outright bans on certain features.
- A protocol for detecting and responding to suicidal ideation and self-harm, usually with referral to crisis services.
- A ban on chatbots presenting themselves as licensed professionals.
- Increasingly, public reporting on how those protocols work.
An operator that builds to this common core is building toward where the patchwork is converging (Troutman Pepper Locke, Orrick).
The federal backdrop
At the October 2026 review, no federal AI companion chatbot statute had been enacted. The main federal action is the FTC's September 2025 Section 6(b) study, which ordered seven companies (Alphabet, Character Technologies, Instagram, Meta, OpenAI, Snap and xAI) to explain how they test and monitor their chatbots' effects, especially on children. A 6(b) study gathers information; it doesn't create binding rules, and no public report had been issued as of October 2026. In Congress, Senator Hawley's GUARD Act, which would ban AI companions for minors, cleared the Senate Judiciary Committee unanimously on April 30, 2026 but had not passed either chamber at that review.
What this means if you operate one of these products
- Build to the strictest law you're subject to. Separate versions per state usually cost more than meeting the highest bar once.
- Plan for January 1, 2027. Oregon and Washington both start then, and New York's signed S9051B also sets that effective date. Both Oregon and Washington give consumers a way to sue. California's SB 1119 follows on July 1, 2027.
- Don't treat "no chatbot law" as "no exposure." Kentucky, Pennsylvania and Texas acted under existing law.
- Re-check at least quarterly. With more than a hundred bills this year, this map changes fast.
Open the switching cost calculator
Updated November 10, 2026: replaced the pending-signature description of S9051B with its October 28 signing and January 1, 2027 effective date, checked against the official Senate record. FPF counts remain a September snapshot; other jurisdictions and the federal section retain their October review scope.
Updated October 2026: added California SB 1119 (signed September 10, 2026), the GUARD Act's committee vote, and FPF's September counts (124 bills, 13 laws signed in 2026; previously 126 bills and "about 15" states). Updated September 2026: an earlier version said three states had a companion chatbot law in effect and relied on older bill counts; it now reflects laws signed in 2026 and current counts. Legislation in this area changes quickly; verify current statutory text and effective dates before relying on this summary. This is a reference summary, not legal advice.